FGDCC: Fine-Grained Deep Cluster Categorization -- A Framework for Intra-Class Variability Problems in Plant Classification

TL;DR

This work tackles intra-class variability in fine-grained plant classification by introducing Fine-Grained Deep Cluster Categorization (FGDCC), which clusters each parent class separately to form latent sub-classes and trains a hierarchical classifier to predict both species and sub-class labels. The approach combines Vision Transformer embeddings with a bottleneck autoencoder to enable class-wise K-Means clustering and end-to-end optimization, including mechanisms to avoid trivial clustering and manage multiple sub-class models. On PlantNet300K, FGDCC achieves state-of-the-art performance while revealing that learned sub-class features may not strongly transfer to the parent classifier, suggesting future work to promote cross-level information sharing. The method demonstrates the viability of class-wise deep clustering for FGVC and provides a platform for further improvements in hierarchical, fine-grained feature learning in long-tailed plant datasets.

Abstract

Intra-class variability is given according to the significance in the degree of dissimilarity between images within a class. In that sense, depending on its intensity, intra-class variability can hinder the learning process for DL models, specially when such classes are also underrepresented, which is a very common scenario in Fine-Grained Visual Categorization (FGVC) tasks. This paper proposes a novel method that aims at leveraging classification performance in FGVC tasks by learning fine-grained features via classification of class-wise cluster assignments. Our goal is to apply clustering over each class individually, which can allow to discover pseudo-labels that encodes a latent degree of similarity between images. In turn, those labels can be employed in a hierarchical classification process that allows to learn more fine-grained visual features and thereby mitigating intra-class variability issues. Initial experiments over the PlantNet300k enabled to shed light upon several key points in which future work will have to be developed in order to find more conclusive evidence regarding the effectiveness of our method. Our method still achieves state-of-the-art performance on the PlantNet300k dataset even though some of its components haven't been shown to be fully optimized. Our code is available at \href{https://github.com/ADAM-UEFS/FGDCC}{https://github.com/ADAM-UEFS/FGDCC}.

Figures (7)

• Figure 1: Example of intra-class variability for the species Smilax aspera L.
• Figure 2: An overview of the Fine-Grained Deep Cluster Categorization Framework (FGDCC).
• Figure 3: A detailed view on the FGDCC architecture.
• Figure 4: Species classification training and validation statistics over the epochs.
• Figure 5: System sub-class classification behavior per epoch - average across the best K values per species.